Table of Contents

• 1  Setup
• 2  Dataset visualization
  • 2.1  Example image/depth and xy-projection of the camera trajectory
  • 2.2  Pointcloud and 3D camera trajectory
  • 2.3  Photometric Image alignment
  • 2.4  Pose error

The autoreload extension is already loaded. To reload it, use:
  %reload_ext autoreload

This repo contains a baic SE3 implementation based on pytorch and a notebook for visualizing the TartanAir dataset

Setup

Download the "carwelding" example trajectory from https://theairlab.org/tartanair-dataset/

Dataset visualization

Example image/depth and xy-projection of the camera trajectory

<function __main__.plot_img(i)>

Pointcloud and 3D camera trajectory

Photometric Image alignment

A soruce image $I_{\text{src}}$ can be warped to match a target image $I_{\text{target}}$ given the pixel-wise depth $d_{\text{target}}$ and relative pose $\Delta T$

Pose error

The alignment error $\|I_{\text{src}} - I_{\text{target}}\|^2$ can be used to judge the accuracy of the depth and pose. The below plot shows how the alignment error changes when we introduce an error in each degree of freedom of $\Delta T$